This is a tiny data set of the first 200 records (40 random datasets and hyper-parameter combinations at 5 epochs) from the complete dataset of 100,000 rows and for demonstration purpose only.
| file | description |
|---|---|
| index.txt | values for the hyper-parameters and the random dataset used for each run and summary statistics. There are 5 rows per parameter combination correponding to 5 different epochs (25, 50, 100, 200, 400) |
| paramInfo.txt | information about each column of index.txt (hyper-parameters + output statistics) |
| runs/[i].txt | one file for each run: predicted labels the points at the i'th run |
| images/[i].png | one plot for each run output: predicted classification. training and test points are represented with white and black stroke respectively. |
Note that unlike the other four datasets, there is no input.txt for this dataset since, for each run the dataset is also randomly generated.
The following images are the classification output plot for the random dataset and hyper-parameters at different epochs.
| epoch: 25 | epoch: 50 | epoch: 100 | epoch: 200 | epoch: 400 |
|---|---|---|---|---|
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |